Date: Wed, 11 Mar 2009 10:30:38 +0200 From: Alexander Motin <mav@FreeBSD.org> To: Yoshifumi SASAKI <sasaki@ongs.co.jp> Cc: freebsd-current@freebsd.org Subject: Re: Try report: Sony VAIO Type P Message-ID: <49B776AE.6050608@FreeBSD.org> In-Reply-To: <49B72F7A.3020709@ongs.co.jp> References: <1236687785.00084874.1236675603@10.7.7.3> <49B6AA99.4090808@FreeBSD.org> <49B6B196.3000305@FreeBSD.org> <49B72F7A.3020709@ongs.co.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
Yoshifumi SASAKI wrote: > Alexander Motin wrote: >> Alexander Motin wrote: >>> Yoshifumi SASAKI wrote: >>>> Sound... succeed to load snd_hda.ko >>>> >>>> hdac0: <Intel SCH High Definition Audio Controller> mem >>>> 0x942c0000-0x942c3fff irq 11 at device 27.0 on pci0 >>>> hdac0: HDA Driver Revision: 20090215_0128 >>>> hdac0: [ITHREAD] >>>> hdac0: HDA Codec #0: Realtek ALC262 >>>> pcm0: <HDA Realtek ALC262 PCM #0 Analog> at cad 0 nid 1 on hdac0 >>>> >>>> I receive many messages.And I try to play MP3 data,But not sound. >>>> >>>> interrupt storm detected on "irq5:"; throttling interrupt source >>>> interrupt storm detected on "irq5:"; throttling interrupt source >>>> interrupt storm detected on "irq5:"; throttling interrupt source >>>> interrupt storm detected on "irq5:"; throttling interrupt source >>> >>> snd_hda supports several interrupt modes: MSI, usual interrupts and >>> polling (interruptless operation). The version you have tested uses >>> MSI by default. You can switch it to use usual interrupts by setting >>> hint.hdac.0.msi=0 in loader.conf, or to use polling by >>> dev.hdac.0.polling=1 sysctl. Also you can try latest CURRENT, as >>> there were some MSI related changes recently. >> >> According to Intel's official SCH chip errata: "High Definition Audio >> Does Not Send Interrupts to the CPU When Using the MSI". So may be >> disabling it is the right way out. >> > > First, I set hint.hdac.0.msi=0 in loader.conf. interrupt storm message > does not stop and not work. > Next, I unset hint.hdac.0.msi=0 and setting dev.hdac.0.polling=1 by > sysctl command. > interrupt storm message stop and sound music:) > > Thanks! You are welcome, but it's just workaround, not a real solution. What's strange to me is that system reports irq11 for hdac0, but it causes irq5 storm, and ndis0 reported as irq11, but causes irq10 storm. I would say there is something wrong with irq allocation. -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?49B776AE.6050608>